Image analysis – Image segmentation – Distinguishing text from other regions
Reexamination Certificate
1999-03-18
2004-05-25
Mehta, Bhavesh M. (Department: 2625)
Image analysis
Image segmentation
Distinguishing text from other regions
C707S793000, C382S305000
Reexamination Certificate
active
06741737
ABSTRACT:
CROSS-REFERENCE TO RELATED INVENTIONS
Not Applicable
STATEMENT REGARDING FEDERALLY FUNDED RESEARCH
Not Applicable
REFERENCE TO A MICROFICHE INDEX
Not Applicable
COPYRIGHT NOTICE
Copyright 1999. A portion of the disclosure of this patent document contains materials which are subject to copyright protection. The owner has no objection to the facsimile reproduction by anyone of the patent document or patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all rights, copyright rights whatsoever.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to the art of extracting information from digital images of forms containing handwritten entries with the images represented as arrays of pixels, and more specifically, to the art of segmenting such images in order to simplify the extraction of handwritten information from the images while ensuring the security of the textual information contained in the images when entering the data into a database. The system of the present invention is particularly useful for segmenting images of standard forms or reports which contains handwritten text and entering that text into appropriate fields of a database.
2. Discussion of Background Information
Transferring data, in particular hand written data, from paper forms containing collected information to data fields within a relational database is a difficult task. In some environments, documents are electronically scanned to generate image files for the documents. Workstations are then provided where one screen or a portion of a screen is used to display all or a portion of a document image and a second screen or a portion of the document display screen is used to provide an input field for a relational database field. The user at the workstation then reads the information from the document image screen and types the information into the input field which populates a data field within a relational database. When documents contain confidential or sensitive information, such as police records, the workstations must be maintained in a secure environment. Nevertheless, because complete documents are presented to a user at a workstation, sensitive information may be viewed by a worker and then disclosed to others, such as newspaper reporters or other unauthorized persons.
Other systems are available to scan documents into a digital image, segment the image into blocks of textual and non-textual data then perform optical character recognition on the text. U.S. Pat. No. 5,430,808 entitled “Image Segmenting Apparatus and Methods” issued Jul. 4, 1995. is an example of such a system. That system segments the image of a scanned document into character or text rectangles and empty rectangles. The analysis of a scanned document image is primarily used to identify the empty rectangles which are then used as a cover set to eliminate the areas of a document that do not contain textual data. That system identifies all of the blank areas by rectangles and then uses that as a cover set to identify the portion of the document image that contains character data which may then be provided to a text extractor for storage. One drawback with this system is that all of the text from the document is complete and available for viewing leading to a possible security breach. Another drawback is the textual data that is generated is not yet in a form suitable for populating the fields of a database. A further drawback is that such as system using optical character recognition does not lend itself to working accurately with information written in widely varying styles. Further, the digital text that is created by an optical character reader is often inaccurate due to skewing or speckling of the original image requiring such text to be manually reviewed. Also, with an automated reader it is not always possible to know where the.error will occur requiring that the entire content of the digital text be compared against that in the original image. It would be advantageous for purposes of security to be able to transmit some of those document portions to one remote station and other document portions to another data entry station for entry by an operator into a relational database field. It would also be advantageous to be able to associate the scanned image with the data fields of database allowing for rapid searching capabilities while retaining the ability to view the entire scanned document if needed.
SUMMARY OF THE INVENTION
The present invention is a system and method that facilitates the entry of data into a relational database from paper documents and permits workers to operate in unsecured areas without jeopardizing the security of the information. Specifically, the system of the present invention electronically scans documents containing textual information that is in most cases handwritten. Such documents include police reports and insurance claim reports. A document identifier is created for each image and comprises a document type identifier and for the document tracking number or serial number. A database record is also created for the document, the record having data fields corresponding to the areas of the image containing the relevant textual information. A segmentation template is defined for each document type or form that is used in the system. The template identifies the various portions of the document containing specific data that is to be entered into the associated database record. Each image area is associated with a particular database field, database record and the original image that the image clip was from. Thereafter, as documents are scanned and received at a processing center, the system identifies the document type and the various fields contained within the document. Image clips are created using a segmentation template corresponding to the database fields used with that document type. The image clips and associated database and image identifiers and a data entry prompt for the particular data field involved are packetized for transmission to a remotely located data entry operator. In addition to a data entry prompt, rules and rule parameters can be provided to assist the operator with the entry of the data. In one embodiment, clips of areas containing the same type of data from a plurality of document images are accumulated, organized into a file, compressed, encrypted, and sent to a remote station over an open network such as the Internet. In other embodiment this process can also be employed with individual packets. The image clips from each individual image are sent to multiple remote workers for processing.
At the remote station, software decompresses and decrypts either the data packet or the data file to generate a display of the image clip and the data entry prompt. The data entry operator reads the image clip and then types or keys in the information into the displayed database field. This data along with the appropriate identifiers is then packetized, compressed and encrypted, and returned to the central site. Alternatively, the processed image clips may be collected into a data file for transmission to the central processing location. At the central site, the data in the data fields received from a remote operator are collected for all portions of each document that was segmented and distributed to multiple workers. In this manner, data are collected for each field of a database record associated with the original image and used to populate a relational database corresponding to that document. Because multiple workers receive information from a single document, workers do not know the identity of the persons involved, or if they know the identity of the person, do not know relevant facts from the report to be able to have a complete picture of the content of the document. This system permits data entry workers to work at unsecured sites without jeopardizing the security of information in a scanned document image.
In alternate embodiments of the present invention, time stamping and continuous real ti
Azarian Seyed
ChoicePoint Asset Company
King & Spalding LLP
Mehta Bhavesh M.
Petty, Esq. W. Scott
LandOfFree
System and method for the secure data entry from document... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for the secure data entry from document..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for the secure data entry from document... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3216073